I don't have any modding or arena experience, and also I'm not a statistician, and I may be a smelly moron, but I hope I can contribute to the analysis meaningfully.
There are two possible ways off the top of my head fortress mode females would show a trend for pain tolerance, and I couldn't tell you for sure without having some source code.
*Females are more likely to spawn with high endurance/willpower/recuperation
*Females with the same endurance/willpower/recuperation stats feel less pain from the same wounds in the same place, due to a gender-based modifier or parameter.
To see if it's the first one, you can wait until your population reaches 200 (or 1000 if you really want to see) and just view every single unit, recording name, gender, skills, attributes. If you use a sql database, you can select the females and count high (willpower or endurance or recuperation) and the low (willpower or endurance or recuperation) and do the same for the males. You can get other reports with a full database of fortress dwarves too, so you could, for example, see if one gender has a high willpower and the other has a high endurance, overall. This might be spoiled by the way you train recruits or assign labor, so you should record migrants and babies as soon as they come in.
For the second case, the best black-box method I could suggest right now would be to have 100 males with the same skills and stats step on a rock trap in the arena (can you do that?) and of the wounded, note how many are in pain, unconscious, or extreme pain. Repeat for females who have the same stat and skill numbers as the males.